Markus Scherer
Markus Scherer
> > Is there any reason not to make the more conservative change? > > I would just like to avoid making any change at all before we find someone...
> There is a pretty long list of differences between what is exported "the old way" and "the new way" Mihai and I are looking at the diffs. Some of...
> Added another commit after Mihai and I reviewed the changes in symbol exports. > > If this passes CI I broke CI. I should have known better: The three...
It looks like the MSVC builds, and thus CI checks overall, are happy now. So we are back to my suggestion: > If this passes CI and no one has...
> If this passes CI and no one has ideas for what else to test and what more to change, then I suggest we remove the first commit (the temporary...
> For the first step here, I'd very much like to use the well-tested and type-safe standard library `std::variant`, as that means less new code that needs to be written,...
If there is an explicit overload for a `const char16_t *`, as in the constructor, then the compiler will prefer that as a perfect match. If there is no raw-pointer...
@roubert ready to un-draft & merge?
FYI: Labels are useful, but for "invariant" we should look at the area field in the Unicode Tools dashboard. I just added your latest two issues there and set the...
We should not need fake UCD 13.1 files for processing *emoji* 13.1. I think we should not need a fake 13.1 value for UCD property Age either. We should revisit...